8

C
H
A
P
T
E
R

8
Syllabus outcomes
5.2.1 Describes and applies problem-solving processes when
creating solutions.
5.2.2 Designs, produces and evaluates appropriate solutions to a
range of challenging problems.
5.2.3 Critically analyses decision-making processes in a range of
information and software solutions.
Overview
This chapter examines articial intelligence and simulations. First, you
gain a basic understanding of articial intelligence and the areas of
articial intelligence, such as expert systems and neural networks.
Second, you will learn the requirements, purposes, advantages and
limitations of models and simulations. The chapter concludes by
discussing the use of modelling and simulation programs.
Artificial
intelligence,
simulation
and
modelling
Artificial
intelligence,
simulation
and
modelling
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 165
166 HI Tech: I nf or mati on and Sof twar e Technol ogy
8.1 Artificial intelligence
Articial intelligence (AI) aims to give computers the ability to think
like human beings. It is the behaviour of a computer that, if
exhibited by a person, would be called intelligent. Some people
argue computers will never have intelligence. They reason a
computer is an electronic tool that can only process data very
quickly but cannot actually think up any new ideas, it does not have
any intelligence. For a computer to be called intelligent, its
behaviours and capabilities must be compared to our own ways of
doing things and thinking.
Consider a simple problem that requires you to choose a piece
of fruita banana, an orange or a pineapple. In making your
decision let us assume that you cannot see, touch or taste the fruit,
but are told the piece of fruit is a sphere. Immediately you know the
fruit must be an orange, but how did you make this deduction? You
have seen many bananas, oranges and pineapples, and from this
experience you know that the oranges are a sphere and that the
other fruits are not. Your intelligent response is probably not based
on having ever been told an orange is a sphere, but from your own
observations. Could a computer solve this simple problem? The
computer is faced with two problems: understanding human or
natural language and recognising the difference between bananas,
oranges and pineapples. It has not learned from observations and
experience that oranges are spheres.
This simple problem raises some interesting questions. Is
choosing a piece of fruit intelligent behaviour? The computer does
not nd this task easy. However, it can perform arithmetic
operations many times faster than any person. What is intelligence?
It is difcult to dene and understand intelligence. People use the
word intelligent to describe somebody who can solve difcult
problems. Intelligence might also be described as the ability to
reason or the power of thought. The Macquarie Dictionary denes
intelligence as the ability to understand, learn and to control
behaviour in any new event. Is a computer intelligent? AI research
has produced some impressive results. Computers can
communicate in human languages, computers have been
programmed to be experts in a range of elds and a computer can
beat the best human in chess. However, most people would not
describe a computer as intelligent.
Articial
intelligence began
in the 1950s with
attempts to write
programs to
challenge human
intelligence, such
as playing chess.
The world chess
champion Garry
Kasparov was
beaten by a
program in 1997.
Lit 1.1
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 166
167 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
Historical perspective
Articial intelligence became available with the development
of the rst electronic digital computer called the Colossus in
1943. It was built by a team led by Alan Turing to break
German military codes (see Figure 8.1). After building this
computer Turing wrote a paper titled Computing Machinery
and Intelligence in 1950. It effectively started the eld of
articial intelligence. The paper contained a concrete test for
determining whether a machine was intelligent. The Turing
test involves two people at the computer. One person sits
alone in a room and types questions into a terminal. The
questions are on any topic. The answers to the questions
appear on the terminal and the person has to determine whether
the answer is written by another person or the computer. If the
person repeatedly chooses the computer then it has demonstrated
intelligent behaviour.
The term articial intelligence was rst used at a conference in
Dartmouth, USA in 1956. The conference was organised by John
McCarthy who is regarded as the father of AI. In the past fty years
the progress in articial intelligence has been steady rather than
spectacular. It includes the following developments:
John McCarthy headed a team that developed the programming
language called LISP (LISt Processing) in 1958. LISP was
completely different from the popular procedural languages
such as FORTRAN. Every expression is a list of calls to
functions. LISP dominated AI research for many years and is
still used today.
Joseph Weizenbaum wrote a program called Eliza (Eliza
Doolittle) in the 1960s. The program acted as a therapist and
encourages the user to talk. For example, if the user types My
father never liked me, the program, recognising the keyword of
father, would respond with Tell me more about your family. If
the user did not type any keywords the program would respond
with a neutral sentence such as Thats very interesting.
Weizenbaum was surprised that people took his program
seriously and thought the computer intelligent. He spent several
years arguing against articial intelligence.
PROLOG (PROgramming in LOGic) was developed in 1970 at
the University of Marseille in France by Alain Colmerauer and
Philippe Roussel. The decision by Japanese computer designers
to work with PROLOG instead of LISP made it very popular.
Figure 8.1 Alan Turing
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 167
168 HI Tech: I nf or mati on and Sof twar e Technol ogy
Many of the features that were once unique to PROLOG are
used in modern object-oriented programming such as Java.
Expert systems rst appeared in the 1970s. They could predict
the probability of a solution under certain conditions. Expert
systems were introduced to a range of applications. MYCIN was
a famous expert system developed at Stanford University. It was
designed to assist non-specialist doctors in the diagnosis and
treatment of bacterial blood infections.
David Marr outlined new theories on articial vision in the
1970s. For example, how is it possible to distinguish an image
based on the shading of an image, basic information on shapes,
colour, edges and texture?
Articial intelligence started to gain pace in the 1980s. Expert
systems were in demand because of their efciency. Large
organisations such as Boeing and General Motors relied heavily
on expert systems. Neural networks were considered possible
ways of achieving articial intelligence.
The worlds rst intelligent robot eye was created in 1990 by
Fujitsu Laboratories. It is capable of detecting high-speed
motion by following a similar procedure to the human brain.
Requirements of artificial intelligence
Articial intelligence requires more processing power and storage
capacity than most other applications. Processing speed and
memory are important requirements for articial intelligence.
Intelligent computers also require a range of input devices to
receive data from the environment. For example, images are
collected using cameras, audio using microphones and data from
the environment using sensors. Sensors are used to detect different
physical conditions such as temperature, light, position and touch.
Output devices for intelligent computers include conventional
devices such as screens and printers, as well as speech synthesisers
and a range of robotic devices to provide movement.
Robots are often used in articial intelligence. They are built to
carry out tasks similar to the tasks carried out by people. Robotic
arms are used to make a response involving movement. They are
programmable and generally consist of a trunk, joints and end
effectors. Some robotic joints only allow linear movement within
one plane and are called prismatic, while other joints enable
rotation and are called revolute. End effectors are tools attached to
the end of the robotic arm to perform the required tasks.
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 168
169 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
Articial vision
enables computer-
controlled devices to see
through the use of
cameras, scanners and
sensors (see Figure 8.2).
Visual images are
captured by a camera and
digitised by an analog-to-
digital converter. These
images are stored on the
computer as a grid of
individual dots called
pixels. Articial
intelligence techniques are used to analyse and interpret the image.
In addition to robots that are equipped with cameras and sensors,
articial vision is used for security screening. A camera can digitise
the image of a persons face or ngerprint to decide entry into a
security area.
Articial intelligence needs software to receive data from input
devices, to process and store this data, and then to make logical
decisions to produce an intelligent response. Software for articial
intelligence is written using non-procedural languages. Non-
procedural languages are used in which the programmer species
what to do and the system determines how to do it. The two most
popular non-procedural languages are LISP and PROLOG.
Exercise 8.1
1 Which person was responsible for this development in AI?
a Outlined new theories on articial vision in the 1970s.
b Developed the programming language called LISP.
c Started the eld of articial intelligence.
d Developed the programming language called PROLOG.
2 Copy and complete the following sentences:
a _______ aims to give computers the ability to think like human
beings.
b MYCIN was a famous_______ system developed at Stanford
University.
c _______ allow intelligent systems to make a response involving
movement.
Figure 8.2 Sensors are used in many robots so that they can see
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 169
170 HI Tech: I nf or mati on and Sof twar e Technol ogy
3 True or false?
a Most people describe a computer as intelligent.
b John McCarthy is regarded as the father of AI.
c John Weizenbaum spent several years arguing for articial
intelligence.
d LISP and PROLOG are two procedural languages.
4 a What is the meaning of the word intelligence?
b Describe the Turing test?
c What is articial vision?
d Why does articial intelligence need software?
Development
5 Some people argue that computers are capable of making
intelligent responses. Other people argue that computers are
machines that follow instructions and will never be intelligent.
Discuss this issue. Your answer should contain references to current
developments in AI research.
6 Construct a table to compare and contrast the intelligence of a
person with that of a computer. What are the tasks performed well
by a computer? What are the tasks people nd easy but computers
nd difcult? Your answer should contain at least ten different tasks.
8.2 Areas of artificial
intelligence
Areas of articial intelligence include intelligent systems, expert
systems and knowledge bases, neural networks, agents and demons.
Intelligent systems
Intelligent systems are systems that receive data from the
environment, react to that data, and produce an intelligent
response. People are intelligent systems. They receive data through
sight, sound, touch, taste and smell, and then use reasoning,
knowledge and feelings to produce speech and movement.
Information technology is being used to create intelligent systems.
One example of an intelligent system is the interpretation of a
natural language.
Natural languages are the languages of humans such as English,
French, Chinese and Arabic. If intelligent systems are going to
produce what people would consider an intelligent response, they
should be able to use natural language. Natural language is the means
by which we share our knowledge. It is not an easy task to teach a
person or a computer a natural language. The main problems are
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 170
171 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
syntax (the rules governing the way in which the words are arranged),
and understanding context to determine the meaning of a word. To
interpret even simple phrases requires a vast amount of knowledge.
For example, teaching the computer the word square must take
into account the following contexts:
this shape is a square
fair and square
the square of two is four
a square meal
out of square
square up
square root
square a debt
youre a square
Natural language processing
is applied in many applications,
however, the results must
be checked by the user. Word
processors contain spell
checkers, grammar checkers
and auto-correct features that
identify and correct likely
spelling errors. Voice
recognition programs provide
96 per cent accuracy even for
continuous speech (see Figure 8.3). Automatic translation
programs are able to accurately convert data to another language.
All these applications have been inuenced by AI research. But the
development of natural languages is one of the challenges facing
articial intelligence. It is going to require further developments in
information technology before a satisfactory solution is found.
Expert systems and knowledge bases
Expert systems provide information and solve problems that would
otherwise require a person experienced in that eld (expert). They
are used in applications such as medical diagnosis, investment
analysis, equipment repair and training. Expert systems provide
information based on the knowledge of the expert. However, their
conclusions are not guaranteed. It is up to the user to accept or
reject the information. The main advantage of expert systems is
their low cost compared with the expense of paying an expert or
team of specialists.
Figure 8.3 Natural language processing
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 171
172 HI Tech: I nf or mati on and Sof twar e Technol ogy
Expert systems ask users a set of questions and compare their
answers to a knowledge base. The knowledge base is a set of general
facts and ifthen rules. If the condition is true, then a certain
deduction is made. For example, if it has four legs, then it is a dog.
However, this deduction is not always valid. A more accurate
deduction is obtained by using a set of ifthen rules. For example, if
it has four legs and if it has a tail and if it barks, then it is a dog.
The inference engine is the part of the expert system that carries
out the reasoning by using the facts, assumptions, theories and
rules. It carries out the reasoning by following a set of strict logical
processes as opposed to the richness of human reasoning. The
inference engine applies the ifthen rules in the knowledge base to
decide what question to ask next. There are two types of inference
engines:
forward chaining is an inference strategy where the user supplies
all the data before the question is asked, or inference made
backward chaining starts with one or more possible solutions
and searches back through the system to determine the
questions to be asked.
A knowledge base is often constructed using expert system shells,
or simply shells. Expert system shells are a ready-made expert
system except they contain no knowledge. When the knowledge is
entered it results in an expert system. Shells provide an interface to
assist the user in creating an expert system. The shell queries the
user for facts and links between the facts. It enters this data into
the knowledge base. The majority of shells represent knowledge
using the ifthen rules. The shell displays the word if on the
screen and the user
enters the conditions.
The deductions are
added after the
conditions. When a set of
ifthen rules are
completed the shell
builds a knowledge base.
Expert system shells
allow expert systems to
be built very quickly. The
great majority of expert
systems in Australia are
built using shells (see
Figure 8.4).
Figure 8.4 An expert system
Tutorial
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 172
173 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
Neural networks
A neural network is a computer system that works like the human
brain and is capable of learning. It contains a large number of
processors connected in a similar way to nerve cells in the human
brain. These processors or nodes form a network and act like cells
in the brain. A neural network nds relationships between sets of
data to develop an understanding of the situation. In an early
demonstration researchers trained a neural network to read aloud.
It learned to do this task by being shown thousands of examples.
Neural networks are very good at nding a link between the data
and a result. They are being used by:
nancial rms for economic forecasting
insurers to decide if a potential customer is a good risk
marketers to predict which products will sell
manufacturers to predict how much material they will need
investment companies to make predictions.
Even though neural networks have been very successful, their
applications are still limited. Most neural networks contain a few
thousand nodes compared to about 100 billion in the human brain.
As a result, neural networks take a very long time to train;
recognising visual images requires the processing of large amounts
of data. Neural networks are in the developmental stages and
billions of dollars are being spent on research.
Agents and demons
Agents are pieces of software designed to search through databases
for relevant data. They use the neural networks to efciently search
vast amounts of data. Agents are independent and can be launched
into a computer system or network to operate in the background.
The most common use of these agents is on the Internet, such as
an email agent or a news agent. An email agent can screen a users
incoming email for those requiring immediate action. Similarly a
news agent is trained to scan articles in a news service and to
deliver a personal bulletin based on a users preferences. Agents aim
to carry out a task faster and more effectively than a human.
Demons are programs that spring to life when they are relevant,
similar to how knowledge springs into your head when it is
appropriate. Demons are an important feature in articial
intelligence as they allow computers to be self-modifying. This
means that computers can teach themselves through experience.
For example, if a computer observes an orange, a demon would be
activated to learn from this experience.
A neural network
was rst proposed
in 1947 by doctors
W. McCulloch and
W. Pitts.
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 173
174 HI Tech: I nf or mati on and Sof twar e Technol ogy
Exercise 8.2
1 Explain the difference between:
a a knowledge base and an inference engine
b an expert system and a neural network.
2 What am I?
a A piece of software to search through databases for relevant
data.
b A computer system that works like the human brain and is
capable of learning.
c A program that springs to life when it is relevant.
d A ready-made expert system except that it contains no
knowledge.
3 Copy and complete the following by replacing the letter in brackets
with a suitable term:
(a) systems ask users a set of questions and compare their answers
to a (b). The knowledge base is a set of general facts and (c) rules.
The (d) applies the ifthen rules in the knowledge base.
4 a What is an intelligent system?
b Why is it difcult to teach a computer a natural language?
c List some of the applications of expert systems.
d How does a neural network work?
Development
5 Computers are becoming more intelligent, not allowing students
to think. Do you agree with this statement? Give reasons for your
answer. Your answer should contain examples of tasks performed
by a computer that could be classied as intelligent.
6 There are many websites that use articial intelligence, such as
predicting share prices and a home-loan analyser. Investigate two
websites that are applying AI. Write a report that describes the
features of these two websites.
7 Use an expert system shell to create an expert system. You will
need to decide on the subject area. Possible subject areas include
choosing a sport, pet, computer or movie. Test your expert system.
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 174
175 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
8.3 Modelling and
simulations
A model is a representation of some
aspect of the real world and a simulation
is the use of that model. Modelling and
simulation are used to make predictions
and examine decisions concerning real
situations. There are many different types
of models used for simulations, including
physical models, mathematical models
and computer models. For example, a
game of Monopoly uses physical models
of streets, money and houses to simulate
the buying and selling of real estate (see
Figure 8.5). However, since Monopoly is a
game that depends on chance, the
simulation is not used to make decisions
concerning the real world.
People establish theories based on what they observe and
measure in real life. Models are then built to test these theories to
see if they are correct. If the model works it can then be used for
simulation. Consider a situation in which temperature needs to be
converted from Fahrenheit to Celsius. A theory is rst proposed
that a formula could be used to do the conversion. A model can
now be developed by measuring a range of Fahrenheit and Celsius
temperatures and determining their relationship. The following
formula is a mathematical model used to test the theory:
C = 5/9(F 32)
C = temperature in Celsius
F = temperature in Fahrenheit
When the model has been tested using the range of Fahrenheit
and Celsius temperatures and the results obtained are correct, it can
be used in a simulation. The simulation uses the model (in this case
the formula) to determine the Celsius temperature given the
Fahrenheit temperature.
Purposes of models and simulations
Modelling and simulation are used when the real situation is time-
consuming, expensive or impossible to reproduce. They are applied
in situations where a prediction is required, or as learning tools.
The word
simulation comes
from the Latin
word similis
meaning like.
Figure 8.5 Game board for Monopoly
Interactive
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 175
176 HI Tech: I nf or mati on and Sof twar e Technol ogy
Many businesses use a spreadsheet to model the way their
business works. The model usually consists of all the income
and expenses of the business, with a simulation to determine its
prot or loss. The data in the model can be changed and the
results seen immediately. It is an invaluable tool and enables the
business to run efciently.
Financial planners create models for different investment
schemes and perform simulations. The amount invested, the
interest rate and the period of the investment can be altered with
each application. When the returns from each scheme are
compared, the nancial planners offer the best investment advice.
Car manufacturers use the computer to simulate car crashes in
order to test new cars and limit the amount of real crash testing.
Engineers are able to watch a model of the car crumple under a
given impact. The cars structure is tested for its reaction to the
stress of a crash at various speeds.
Simulators are devices used for training and in experiments to
simulate movement, ight or some other condition (see Figure
8.6). They are increasingly used in industry where it is too
dangerous or too expensive for people to practise on the job.
Drilling simulators are used in the oil and gas industry to train
operators on the drilling rig oor. The trainee operators read
gauges and dials, set drilling speeds, and use the brakes under
the supervision of an instructor. The instructor can introduce
faults and problems to monitor the progress of the operator.
Simulators have also been used for the space shuttle mission.
Astronauts are trained using a simulator, which is a model of
the shuttle. It is very realistic and allows the astronauts to
experiment and react to the conditions they will be encountering.
Scientists use simulation
in many ways. Molecular
model kits are used to
simulate the shape of
complex molecules and
have led to the discovery
of chemical compounds.
Computers can also create
these models and allow the
viewer to rotate the model
on the screen. This allows
scientists to examine the
shape of compounds
Figure 8.6 A ight simulator
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 176
177 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
before they are synthesised. This can be useful in drug synthesis
where the shape of the molecule is an important factor in its
physiological response.
Astronomers have used a simulation to nd the position of stars
that could not be seen from earth. The model uses data from
radiation entering the earths atmosphere, and the simulation
has enabled a supercomputer at NASA to construct a more
detailed map of the universe. A theoretical model of a black
hole has been used by scientists to simulate its ow of gases.
This simulation involved 25 000 variables, 10 000 steps in each
experiment and a solution consisting of 1.25 billion numbers.
Advantages and limitations
Modelling and simulation provide many advantages. In many
situations it is safer, cheaper and quicker to use a model rather than
the real situation. Designing a model of a car using a computer and
simulating a range of tests is much more efcient than building the
car. Minor changes to the model are easily completed and another
simulation completed. In some situations, such as preparing
astronauts for a space mission, a simulation is the only way of
training. In business, models and simulations are used to ensure
the business is protable and to determine future opportunities.
A simulation is only as good as the model and the data used to
create it. If the model is incorrect or the data entered is inaccurate
then the simulation will produce the wrong result. In addition, if
the model is too simple and does not take into account a major
factor, the result of the simulation will be unreliable. For example,
a model used by nancial advisors that did not take into account
rising interest rates would result in inaccurate advice when the
simulation is performed.
Exercise 8.3
1 What am I?
a The use of a model.
b Software used by many businesses to model the way their
business works.
c A model of the space shuttle used by astronauts for training.
d A type of computer used at NASA to construct a more detailed
map of the universe.
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 177
178 HI Tech: I nf or mati on and Sof twar e Technol ogy
2 Copy and complete the following sentences:
a A _______ is a representation of some aspect of the real world.
b If a model works it can be used for _______.
c Modelling and simulation are used when the _______ is time-
consuming, expensive or impossible to reproduce.
d _______ create models for different investment schemes and
perform simulations.
3 Unjumble these words:
a nsiulaomti
b lodme
4 a How are models and simulation used by car manufacturers?
b How are drilling simulators used by trainees?
c Describe the way scientists use simulation.
d What are the main advantages of modelling and simulation?
Development
5 Most computer games involve a simulation. Compare and contrast
a computer game to the real situation.
6 Modelling and simulation programs are freely available from the
Internet. Use a simulation program and perform the simulation.
Write a report to describe the use of this program. Evaluate the
value of this program.
7 Modelling and simulation make life too predictable and are not
beneting our society. Do you agree with this statement? Why? Do
research to nd current applications of modelling and simulation to
illustrate your answer.
8.4 Requirements of
models and simulations
There is a range of hardware and software that can be used to
construct models and perform simulations.
Hardware
Creating a model and performing a simulation can be completed
on most computer systems. However, to achieve accurate and more
realistic simulations hardware capable of performing computations
quickly and with a high degree of precision is required. If the
simulations are complex, such as scientic applications, then a
supercomputer or mainframe may be required. These computers
can process data at very high speeds and have extensive storage
Lit 1.2
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 178
179 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
capabilities. Other hardware features include simulators, analog
computers and parallel processors.
Simulators are devices used for training and experiments to
simulate movement, ight or some other condition. For example,
ight simulators have become an effective way to train pilots. The
simulator consists of a model of a cockpit with authentic seats,
realistic instrument panels and the ability to simulate movement
using hydraulic legs that are controlled by the computer system.
The simulation presents the trainee with different conditions to test
their skills, such as the weather conditions at various airports or a
mechanical failure in the aircraft. Flight simulators allow pilots to
be trained without risking human life and valuable aircraft.
Simulators are also used to train car drivers, racing drivers, crews of
large ships or controllers of any other vehicle.
An analog computer is a device that processes analog data. It is
designed to perform arithmetic functions on numbers that are
represented by a physical quantity, such as a voltage. Analog
computers can accept data directly from peripheral devices in
analog form without having to rst convert it into digital form.
This allows analog computers to process data at very high speeds
making them suitable for modelling and simulation.
Parallel processing is the use of several processors to share the
calculations of a large task. Parallel processing is much faster than
using one central processing unit to carry out several instructions
one after the other. Parallel processing divides the processing task
between a series of processors. Parallel processing provides high
speed calculations needed in many modelling and simulation
systems.
Virtual reality is an extension of a
simulation program where a person
interacts with the simulated world. It
refers to a computer program and
other devices that enable a person to
perform tasks virtually with all the
function and effectiveness of the real
situation. Images are displayed using
special screens mounted in a helmet,
and the user interacts through the
use of a special glove that allows
them to move through the simulated
world by pointing and grabbing (see
Figure 8.7). Figure 8.7 Virtual reality
Lit 1.3
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 179
The Simula
(SIMUlation
LAnguage) is
a programming
language designed
and built by
Ole-Johan Dahl
and Kristen
Nygaard at the
Norwegian
Computing Center
in Oslo.
180 HI Tech: I nf or mati on and Sof twar e Technol ogy
Software
Software for modelling and simulation systems consists of
modelling programs, programming languages, spreadsheets and
simulation programs.
A modelling program is any program that allows the user to
create models and simulate a real situation. Even though
spreadsheets are used as modelling programs, there is more
sophisticated software that specialises in creating models. These
programs are very complex, usually containing a programming
language, and have more exibility in viewing and reporting data.
One example is called EasyFlow. This modelling program is for
uid dynamics and has applications in aerospace, architecture,
electronics and the environment.
All general purpose programming languages such as BASIC
and C++ are used for modelling and simulation. However, specic
simulation languages, such as Dynamo and Simula, have been
developed for modelling and simulation applications. They
contain denitions of the essential elements of a model and
procedures for modifying and observing the model.
Simulation programs have been developed for many different
elds, such as business, science and mathematics. These programs
contain a model and allow the user to make predications and
decisions based on the model (see Figure 8.8). In business, packages
such as Marketplan and Brandaid are used for marketing, Callplan
and Detailer for sales, and Mediac helps prepare advertising media
schedules. There are also many simulation programs in science and
mathematics that
are used to test and
teach theories. For
example, a program
called Catlab
teaches the
principles of
genetics by
modelling the
breeding of cats
with different
colours.
Figure 8.8 Simulation program
Lit 1.5
Lit 1.4
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 180
181 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
Project: Gold Coast conference
Nita needed to organise a conference on the Gold Coast for her staff.
She had to determine the cost of the conference, an outline of an
itinerary and construct a budget. The problem was solved using the
four stages in project development.
Dene and analyse the problem: A search of the Internet was
completed to gather information on motels, places to visit, travel
arrangements and possible dates. A project plan was written and
some initial ideas discussed with her supervisor.
Design possible solutions: Four possible solutions were examined in
a feasibility study. Nita recommended a three-day conference at
Conrad Jupiters. PowerPoint was used to present the solution to
her supervisor for approval. The supervisor approved the solution if
the cost could be reduced by 10 per cent.
Produce the solution: A proposed budget was developed using a
spreadsheet program. Nita used whatif predictions to obtain the
cost reduction. The venue and air travel were booked. Each
member of staff was advised of the arrangements and their
responsibilities during the conference. The conference was
completed and invoices received.
Evaluate the solution: The cost of the conference was below the
proposed budget. Each staff member completed an evaluation of
the conference. The report on the conference, including the
budget, was submitted to the supervisor for accountability.
Tasks
1 Why was a spreadsheet program used to construct a budget?
2 Investigate any templates that could be used to construct a budget.
Describe the labels, values and formulas in these templates.
Exercise 8.4
1 True or false?
a Analog computers can accept data directly from peripheral
devices in analog form.
b A special screen mounted in a helmet is often used for parallel
processing.
c Programming languages cannot be used for modelling
simulation.
d Simulation programs have been developed for many different
tasks.
2 Copy and complete the following sentences:
a The ability to perform _______ quickly is an important aspect in
modelling and simulation systems.
ICT 8.1
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 181
182 HI Tech: I nf or mati on and Sof twar e Technol ogy
b In virtual reality the user interacts with the simulated world
using a _______.
c If simulations are complex, such as scientic applications, then
_______ may be required.
3 What am I?
a A device used for training and experiments to simulate
movement, ight or some other condition.
b A program that allows the user to create models and simulate
the real situation.
c An extension of a simulation program where a person interacts
with the simulated world.
d It divides the processing task between a series of processors.
4 a List some of the advantages of using a simulator.
b What is an analog computer?
c Describe two specic simulation languages.
d Describe a simulation program.
Development
5 Simulation of natural disasters, such as earthquakes, is an attempt
to save lives at the risk of interfering with nature. Does this
application of technology further separate society from natural
systems and result in a heartless attitude to the environment? Give
reasons for your answer.
6 The requirements of models and simulations are constantly
changing. Do research to examine the hardware needs for
operating simulation programs. Write a summary of your
investigation.
8.5 Using models and
simulation programs
Data in a simulation represents the state of the model at any
particular time. If the data is changed in the model then so is the
simulation. Data is stored in variables and parameters.
Variables are items of data that may change. A variable name is
used to store the data or value of the variable in a location of memory.
For example, if A5 is used as a variable and given the value 10 (A5
= 10), then a specied memory location referred to as A5 will hold
the number 10. In a modelling and simulation program, data that is
stored in variables can be easily changed.
Parameters are variables whose values cause particular instructions
to be completed. During the course of the simulation a parameters
value might be continually changing, the model uses these values
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 182
183 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
to help determine the result of the simulation. For example, in a
modelling and simulation program the variable A5 might also be
parameter. If its value is 10 (A5 = 10) then a set of instructions is
completed. However, if its value is 12 (A5 = 12), a different set of
instructions is completed. The result of the simulation depends on
the parameter A5.
Spreadsheets
A spreadsheet is a software package that can be used for modelling
and simulation. It is a rectangular grid made up of rows and
columns to organise and store data that require some type of
calculation. A spreadsheet creates a model using formulas to
represent the real situation. For example, a spreadsheet could be
used to calculate loan repayments. It would use a formula that
relates the interest rate, time period and present value of the loan.
Spreadsheets can present data in tables and charts and make
predictions based on trends.
All spreadsheets consist of rows that run horizontally across the
screen and columns that run vertically down the screen. Rows are
usually numbered 1, 2, 3, 4, and so on, while columns are often
named in alphabetic terms A, B, C, up to Z; then AA, AB, AC up to
AZ; then BA, BB and so on. The intersection of a row and column in
a spreadsheet is called a cell. The position of the cell in the
spreadsheet is called its cell
address (or cell reference) such
as A2. A range is a group of
adjacent cells in a single row or
column, or in several adjacent
rows and columns. For example,
the range B5:C7 contains the
data in cells B5, B6, B7, C5, C6
and C7.
A spreadsheet is very
particular about the type of data
that is entered into each cell.
There are three main types of
data called labels, values and
formulas (see Figure 8.9):
Label is text entered into
a cell to provide some
explanation of the
spreadsheet. Calculations
Figure 8.9 A spreadsheet containing labels, values and
formulas
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 183
184 HI Tech: I nf or mati on and Sof twar e Technol ogy
are not carried out using this data. Labels are used for headings
in rows and columns.
Values are numbers stored in a spreadsheet on which
calculations can be carried out.
Formulas are instructions to perform a calculation. The answer
to the calculation will always appear in the cell that contains the
formula. The formula itself is not shown. The user can create a
formula or select a predened function such as average.
Formulas often contain cell addresses that are used as
parameters.
Spreadsheets have many features that make them easy to use. Some
of the features include editing data, workbooks, templates, macros
and charts.
Editing data is completed on the spreadsheet or using the
formula bar. Cut and Paste, Copy and Paste, or Drag and Drop
are used to copy values, labels, cell references and formulas into
other cells. The Fill down or Fill right instruction copies the
contents of a cell into a range of adjacent cells.
A workbook is made up of one or more spreadsheets called
worksheets or sheets. These sheets allow the user to organise
related information in a single le.
A template is a document created for repeated use. A
spreadsheet template contains labels and formulas to solve a
particular problem. The user enters the values to obtain the
required solution. A template saves time and effort.
A macro is a series of commands stored in a le that can be
executed by pressing a few keys. Macros allow many operations
to be performed automatically, such as opening menus,
choosing commands or entering text.
A chart is a graphical representation of numerical data. Charts
convert data in rows and columns into a picture that is read at a
glance. Charts make data easy to understand. Trends are
revealed and comparisons can be made instantly. Charts assist
people to make quick and accurate decisions.
Whatif predictions
Whatif predictions are a process of making changes to the data and
observing their effects. It answers whatif? questions: What if the
interest rate increases by 1 per cent, what would be the effect on my
loan repayments? What if I buy a car with a fuel consumption of 8
The word formula
has two plural
formsformulas
and formulae.
The latter is most
often used in
scientic work.
Lit 1.6
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 184
185 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
L/100 km instead of 10 L/100
km? How would this affect
the cost of fuel over the next
two years? Spreadsheets
instantly give the user the
answer to whatif
questions. You change the
data and observe the effect.
Goal Seek is a command
used to make whatif
predictions. If the user knows
the result of a particular
formula, but not the value to
be entered, then Goal Seek
will determine this value. For
example, in Figure 8.10 Goal
Seek is used to determine the
sales of cola in December
that will result in average
monthly sales of eighty for
the past six months.
Spreadsheet design
A spreadsheet must be well designed and easy to use. There is no
correct way to design a spreadsheet. Each particular problem has its
own design requirements. However, a well designed spreadsheet has
four easily identiable areas called the instruction, input,
calculation and output (see Figure 8.11):
Instruction areainformation about the spreadsheet or
directions for use of the spreadsheet. It is usually at the top of
the spreadsheet and includes a title, a description, the author
and date. Larger spreadsheets include a brief outline of their
structure, directions and parameters.
Input arealabels for headings and values on which
calculations are based. The format of values and labels is often
completed after the data has been entered. For example, when
entering money values do not include the $ sign as these values
can be format to currency after the numbers have been entered.
All inputs should be labelled clearly so the user knows where
and how to enter the data.
Figure 8.10 Whatif predictions and Goal Seek
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 185
186 HI Tech: I nf or mati on and Sof twar e Technol ogy
Calculation areaformulas and functions complete the work of
the spreadsheet. The calculation area is the heart of a
spreadsheet.
Output areadisplays the result of the spreadsheet. In many
spreadsheets the output area and the calculation area will be the
same. All outputs should be clearly labelled so the user under-
stands the results. Presentation of data should be suitable to the
type of data and its use. Often a chart will display the information
in more understandable form for a decision to be made.
Exercise 8.5
1 Explain the difference between:
a variables and parameters
b labels and values
c templates and macros
d an instruction area and an input area of a spreadsheet.
2 What am I?
a The instructions in a spreadsheet to perform a calculation.
b One or more worksheets.
c The process of making changes to the data and observing the
effects.
d A graphical representation of numerical data.
3 Copy and complete the following by replacing the letter in brackets
with a suitable term:
All spreadsheets consist of (a) that run horizontally across the
screen and (b) that run vertically down the screen. The (c) of a row
and a column is called a cell. A range is a group of (d) cells in a
single row or column.
Figure 8.11 A well-designed spreadsheet.
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 186
187 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
4 a What is a spreadsheet?
b How are rows and columns numbered in a spreadsheet?
c List the three main types of data in a spreadsheet.
d How is data edited in a spreadsheet?
Development
5 Create the three spreadsheets shown in this section. Answer the
following questions by referring to these spreadsheets.
a What is the parameter? Give an example.
b Use the Fill Down command to enter data. Describe the steps
to use this tool.
c Explain the difference between an absolute and relative
reference. Give an example.
d What is a function? List ten different types of functions and
give an example using correct syntax.
e Use Goal Seek to perform a whatif prediction. Describe the
steps in your prediction.
6 Design a spreadsheet to analyse your time management. For each
day of the week, enter the amount of time (in hours) you spend on
sleep, eating, television, homework, study and other activities. Use
the spreadsheet to total the hours for each activity and nd the
weekly averages. Design the spreadsheet using four easily
identiable areas. Construct appropriate charts from this
spreadsheet.
ICT 8.2
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 187
188 HI Tech: I nf or mati on and Sof twar e Technol ogy
Part A: Multiple choice questions
Select the alternative (a), (b), (c) or (d) that best answers each question.
1 Which of the following is the person
who effectively started the eld of
articial intelligence?
a Joseph Weizenbaum
b John McCarthy
c David Marr
d Alan Turing
2 Which of the following describes
articial intelligence?
a Aims to give computers the ability
to think like human beings
b The ability to understand, learn
and to control behaviour in any
new event
c A computer system that works like
the human brain and is capable of
learning
d The ability to give an intelligent
response using natural languages
3 Which of the following is the set of
general facts and ifthen rules in an
expert system?
a Database
b Knowledge base
c Inference engine
d Intelligent agent
4 Which of the following describes
expert system shells?
a Receive data from the environment,
react to that data and produce an
intelligent response
b Provide information and solve
problems that would otherwise
require a person experienced in
that eld
c Carry out the reasoning by using
the facts, assumptions, theories
and rules
d Software used to construct a
knowledge base in an expert
system
5 Which of the following is a
representation of some aspect of the
real world?
a Model
b Simulator
c Demon
d Agent
6 Which of the following is not an
advantage of modelling and
simulation systems?
a Predicts the result of an event
b Teaching tool to prepare trainees
c Inexpensive to use compared to
the real situation
d Results of the simulation cannot
be changed
7 What is the name of a device used
for training and experiments to
simulate movement, ight or some
other condition?
a Virtual reality
b Simulator
c Experimenter
d Training machine
8 Which of the following is a
simulation language?
a Simula
b Sim City
c Java simulator
d Simlang
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 188
189 8 Ar ti f i ci al i ntel l i gence, si mul ati on and model l i ng
9 Which of the following is an
instruction to perform a calculation
in a spreadsheet?
a Formula
b Value
c Label
d Macro
10 Which of the following is not a
method of editing data in a
spreadsheet?
a Cut and paste
b Drag and drop
c Fill down
d Formula bar
Part B: Matching the term
For each of the following statements (1 to 10), select from the list of terms (a to j)
the one that most closely ts the statement.
Statements
1 Provides information and solves
problems that would otherwise
require a person experienced in that
eld.
2 A system that works like the human
brain and is capable of learning.
3 A set of general facts and ifthen
rules.
4 A representation of some aspect of
the real world.
5 A device used for training and
experiments to simulate movement,
ight or some other condition.
6 Item of data that may change.
7 A rectangular grid made up of rows
and columns to organise and store
data that require some type of
calculation.
8 A process of making changes to the
data and observing their effects.
9 A system that receives data from the
environment, reacts to that data and
produces an intelligent response.
10 It aims to give computers the ability
to think like human beings.
Terms
a Articial intelligence
b Expert system
c Intelligent system
d Knowledge base
e Model
f Neural network
g Simulator
h Spreadsheet
i Variable
j Whatif predictions
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 189
190 HI Tech: I nf or mati on and Sof twar e Technol ogy
Part C: Extended-response questions
Write at least one paragraph for each of the following:
1 Why is it so difcult to dene
articial intelligence? How would
you test whether a computer was
intelligent?
2 Developments in information
technology will eventually allow
computers to create their own
knowledge bases. Explain this
statement.
3 You are a passenger on a plane and
have just learned that it will be
landing automatically using an expert
system. What is an expert system?
How do you feel about this situation?
Give reasons for your answer.
4 Describe a situation that uses a model
and simulation. What are the
advantages of this simulation? Outline
the requirements of this simulation.
5 Describe the requirements of models
and simulations. Divide your answer
into hardware and software.
6 Spreadsheets are often described as
the rst killer application. People
bought personal computers to use
spreadsheets. Why are spreadsheets so
popular? Outline the advantages of
spreadsheets compared to manual
methods of performing the same tasks.
Project: Pizza shop
A pizza shop is having problems meeting its costs. It needs to decide whether it
should increase or decrease its prices. Visit a local pizza shop to obtain current
price structures, types of pizzas and the typical sales during the week. Design a
spreadsheet model that could be used to estimate the effects of different prices
on revenue and the quantity of pizzas purchased. Survey students on their pizza
preferences and use this information to make a recommendation to the owner of
the pizza shop. Demonstrate the use of whatif predictions to assist your decision.
Your solution should be developed using the four stages in project development.
Project 1.1
Project 1.2
Project 1.3
Cross project 1
Cross project 2
Cross project 4
e Tester
313 ISTech Chapt 08.3pp 23/9/04 11:50 AM Page 190